AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

MCP Security Bench (MSB): Benchmarking Attacks Against Model Context Protocol in LLM Agents

Dongsen Zhang, Zekun Li, Xu Luo, Xuannan Liu, Pei Pei Li, Wenjun Xu

ICLR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 39 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
combine PI and UI attacks into a mixed attackLLM agents (avg of 10) · MSBASR: PI 20.21, UI 45.69 -> PI-UI 56.07Improved
use user-impersonating tool response (UI) instead of false error (FE)LLM agents (avg of 10) · MSBASR: 39.21 -> 45.69Improved
add MCIP detection-based defense to MCP agentsLLM agents (avg of 10) · MSBASR: 40.35 -> 28.69Improved
add MCIP defense to MCP agentsLLM agents (avg of 10) · MSBPUA: 56.50 -> 48.91Hurt
add MCIP defense to MCP agentsLLM agents (avg of 10) · MSBNRP: 33.70 -> 34.88No clear effect
add MCIP defense to GPT-5 agentGPT-5 · MSBNRP: 52.99 -> 46.73Hurt
use more capable LLM backbone (higher PUA)LLM agents · MSBASR: DeepSeek-V3.1 ASR 60.94 vs Llama3.1 8B 19.74Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy